Authors - Ravi Tene, Dasari Kalyani, N. Sudhakar Yadav, Kondabala Renuka, Gunupudi Rajesh Kumar, Nimmala Mangathayaru Abstract - A deepfake is a misleading video or image that looks genuine. GANs (Generative Adversarial Networks) are the known name in the domain of machine learning. GANs generate a huge amount of fake human writing with deep-learning-wide-models. The generator model learns to sample points from a latent space so that new samples of the same distribution can be fed in and produce different observable model outputs. Deepfakes for most applications can be convincingly created using Generative Adversarial Networks (GANs). There are fears on the Internet related to deepfake. However, the authors use ResneXt and LSTMs for using Deep Learning Network to identify fake areas of deepfake uses Python facial recognition and C++ visual libraries to identify a face in this video. Fake videos are further validated using models trained on various edge groups.
Authors - Vanishree Pabalkar, Rahul Dhaigude Abstract - The purpose to carry out the research study is to understand the concepts of introducing advanced technology that prevails in EV market and the challenges and strategic solutions. This research validates customer feedback and allows companies to get closer to the true opinion of potential Indian customers. In addition, this can eliminate misunderstandings and problems to trade better. This study was conducted to understand the factors that influence the choice of an electric vehicle. The current research has been conducted to study the purchasing behavior of consumers when purchasing electric cars by identifying the importance ratings assigned to different factors during the selection process. in the electric car, and analyze the reasons for the brand's success by identifying the levels of excellence. Current users use different types. Characteristics and identification of the gap between importance ratings and current ratings.
Authors - Vatsal Suchak, Harmin Rana, Ayush Verma, Nilesh Dubey, Hardikkumar Jayswal, Dipika Damodar, Chirag Patel Abstract - Cattle farming plays a crucial role in global food production, but monitoring the health of large herds poses significant challenges. Traditional manual inspections are inefficient, reactive, and prone to error, highlighting the need for scalable, automated health monitoring systems. This paper introduces a smart cattle health monitoring system that utilize the Internet of Things technology and machine learning algorithms to provide real-time health tracking. The system used a proposed wearable devices equipped with ESP32 microcontrollers and sensors to monitor cattle’s vital parameters, such as body temperature and heart rate. Data collected from the devices is transmitted to a local XAMPP server and analysed by an edge-computing device, Jetson Nano, which processes the data using supervised and unsupervised machine learning models for anomaly detection. If health anomalies are detected, the system sends real-time alerts to farmers, allowing for timely intervention. The system’s design focuses on local processing for low-latency performance, scalability for large herds, and robust security measures. This project demonstrates the potential of IoT-based livestock health monitoring systems to enhance productivity, improve animal welfare, and reduce economic losses due to illness.
Authors - Teena Bambal, Dipesh Chavan, Nikhil Gadiwadd, Deepak M. Shinde Abstract - This survey paper investigates the application of artificial intelligence (AI) and machine learning (ML) techniques for the early detection and diagnosis of liver disease. Traditional methods of liver disease diagnosis, such as blood tests and imaging techniques, can be time-consuming and prone to human error. AI-based approaches offer the potential to improve accuracy, efficiency, and accessibility of liver disease diagnosis. The research investigates a range of AI and ML algorithms, such as decision trees, support vector machines, random forests, neural networks, and deep learning models. These algorithms are applied to analyze large datasets containing patient information and medical test results. The performance of the models is evaluated using metrics such as F1-score, precision, accuracy, recall, and AUC. The findings demonstrate the effectiveness of AI-based approaches in accurately detecting liver disease. Compared to traditional methods, AI models can provide more reliable and timely diagnoses, leading to improved patient outcomes. The research highlights the potential of AI to revolutionize the field of liver disease management and improve global healthcare.
Authors - Yatin Nargotra, Tanya Jagavkar, Tushar Birajdar, L.P.Patil Abstract - The Mahavitaran Help App is a mobile application aimed at revolutionizing the process of reporting electrical outages in India. Current systems for outage reporting are often slow, inefficient, and lack the integration needed to quickly address user complaints. The Mahavitaran Help App simplifies this process by allowing users to submit complaints via mobile devices, integrating location services with Google Maps and supporting the upload of complaint-relevant images. Moreover, this project introduces a critical migration from Firebase to AWS or Google Cloud, offering improved scalability, reliability, and faster processing of complaints. This paper presents a detailed review of existing mobile complaint management systems and explores cloud-based scalability and security features, including OTP authentication for securing user data.
Authors - Rani S. Lande, Amol P. Bhagat, Priti A. Khodke Abstract - Visual memes have become a pervasive form of communication in digital spaces, presenting a unique challenge and opportunity for content analysis due to their blend of visual, textual, and often humorous elements. This paper reviews and synthesizes methodologies employed in the analysis of visual memes, aiming to provide a comprehensive overview of current practices and future directions. The methodologies discussed encompass a range of approaches, including qualitative, quantitative, and mixed-methods strategies. Qualitative methods delve into semiotic analysis, exploring how visual and textual components interact to convey meaning and cultural references. Quantitative approaches employ computational tools to analyze large datasets, focusing on metrics such as image recognition, sentiment analysis, and virality metrics. Mixed-methods studies combine these approaches to offer nuanced insights into the multifaceted nature of visual memes. Challenges in visual memes content analysis include the rapid evolution of meme formats, cultural context sensitivity, and the ethical implications of meme reuse and modification. Additionally, the paper explores emerging trends such as deep learning techniques for image recognition and natural language processing for text analysis within memes. By synthesizing these methodologies, this paper aims to provide researchers and practitioners with a foundational understanding of how to effectively analyze visual memes, highlighting opportunities for interdisciplinary research and applications in fields ranging from communication studies to digital humanities and beyond.
Authors - Nitika Sharma, Rohan Patel, Hardikkumar Jayswal, Nilesh Dubey, Hasti Vakani, Mithil Mistry, Dipika Damodar, Shital Sharma Abstract - This study explores the use of advanced machine learning models to forecast trends in Apple’s stock market performance. Stock market forecasting presents a formidable challenge, given the inherent volatility and unpredictability of market behavior. The study investigates various advanced models, such as Logistic Regression, XGBoost, Artificial Neural Networks, Recurrent Neural Networks, Long Short-Term Memory (LSTM), and ARIMA, for predicting stock prices. Analyzing historical data spanning from 2014 to 2024, which includes Apple's daily stock prices and trading volume metrics, the research applies Grid Search optimization to fine-tune model parameters, thus enhancing predictive accuracy. The findings reveal that LSTM achieved the highest accuracy at 96.50%, followed closely by ARIMA at 90.91%. These results highlight the critical role of machine learning in improving stock price predictions, thereby facilitating more informed investment decisions.
Authors - Anant Nikam, Atharva Gangapure, Samarth Deshpande, Sonali Shinkar Abstract - Among the primary concerns in the digital era, secure sharing of data stands prominent. The integrity, confidence, and authenticity of information being shared form a very significant concern. Blockchain technology promises much towards overcoming such challenges due to its decentralized and immutable nature. It significantly enhances data security through its use of encryption techniques, such as hashing and digital signatures, which eliminate the need for middlemen in transactions while also reducing the probability of data breaches. It leverages smart contracts that provide mechanisms for automating access controls wherein data is shared appropriately, according to agreed terms, among the participants in a trustless environment. Some practical illustrations of use cases from healthcare, supply chain management, and finance are found in the context provided below. Findings thus reveal the needed innovation to revolutionize the state of secure data sharing on blockchain technology by providing a strengthened, decentralized infrastructure that promotes trust, transparency, and accountability of stakeholders involved.
Authors - Dibyendu Rath, Arunangshu Giri, Dipanwita Chakrabarty, Puja Tiwari, Satakshi Chatterjee, Shamba Chatterjee Abstract - The study reveals how mobile health apps and information technology can take a pivotal role for healthcare improvement, especially for rural population, where people suffer from medical infrastructural inadequacy. Healthcare apps facilitate the users by providing a 24x7 accessibility at a cost-effective rate. The study used cross-sectional surveys for analyzing responses across different demographic profile, like age, gender, qualification, income group etc. This study has identified some key factors that help to engage customers with healthcare apps. The study also reveals that trust on healthcare app will enhance intention to adopt healthcare apps and trust will be positively influenced by Perceived Benefits (PB). Again, trust will be negatively induced by Perceived Risks (PR) and Technology Anxiety (TA). Four hypotheses were made to validate the relationships among the factors. Finally, the study balances the benefits and risks of using healthcare apps and guides how m-health technology can increase adoption intentions.
Authors - Vasu Agrawal, Nupur Chaudhari, Tanisha Bharadiya, Manisha Sagade Abstract - The recent progress in AI and deep learning has significantly transformed the public safety landscape, particularly in the area of real-time threat detection in public domains. This comes with increased complexity and density as urban environments become more complex; traditional surveillance systems are no longer enough for monitoring large crowds, detecting potential threats, or ensuring public safety. This has necessitated the development of automated systems that could process large volumes in real time to pick anomalies, suspicious behaviors, and objects liable to imperil security. We delve into the core methodologies that object detection models, such as YOLO, Faster R-CNN, SSD, and compare them .To further improve the accuracy in detection of anomalous and illegal activities and reduce false positives and negatives we created a custom dataset by fusing data from different sources, these systems enhance the overall reliability of the surveillance systems.
